Does HumanizerOps optimize only for a lower detector score?

No. Evaluation also covers semantic drift, protected content, reproducibility, latency, and model provenance.

Can teams inspect each transformation route?

Yes. Profiles are explicit and versioned, and supported translation routes expose their intermediates through the API contract.

RUNS / JUL-14 / RUN-7F31A9

Humaniser engine workbench

live EU API

The canonical detector, editor, profiles, model evidence, and revision workflow in the HumanizerOps skin.

RUN ID: RUN-7F31A9
SOURCE: sha256:b7e9…23af
POLICY / RECIPE: v24 / editorial-safe@7
EXECUTED: 14 Jul 2026 · 18:42:09
COST: $0.1842 / $0.25 max
LATENCY: 12.84s / 20s max
RETENTION: ◆ HASH ONLY

Runtime transparency

Model routes available to this workbench

The scan and run receipts are authoritative: they name the model actually used for that request.

DETECT

Primary detector

desklib/ai-text-detector-v1.01

Document and sentence evidence. Deep scan can add independently reported layers.

REWRITE

Balanced local

humarin / T5-base route

Restrained lexical rewrite with a final detector pass.

REWRITE

Structural variation

dipper / optional 11B route

Availability is reported at runtime; unavailable models are never silently substituted.

TRANSLATE

NLLB round trips

translation-*

Single and multi-pivot routes expose the path and intermediate text in the run trace.

EDITORIAL

Provider-resolved rewrite

claude-rewrite

The returned receipt names the concrete provider model; citations and claims still require review.

A listed route is not a promise that it is loaded or enabled. Each result shows actual availability, model name, route, and missing evidence.

SAFE SELECTED · EDITORIAL REVIEW RECOMMENDED

Atlas Edit / atlas-3.2-long

Candidate C-017 passes every hard constraint. One detector timed out, so the result is selected on available evidence—not declared best overall.

SEMANTIC: 0.947
DETECTOR SIGNAL: 0.182
HARD GATES: 8/8
REPORT: STATIC P0

Missing evidence: Detector Northstar-1 timed out at 8.0s. No score was imputed.

1safe selectedof 3 candidates

8/8hard constraintssource protected

1detector reviewsignal is secondary

1missing evidencetimeout retained

14mfreshnessprovider registry

GUIDED INTAKE

Replay benchmark

✓ READY

MAX CANDIDATES REQUEST API KEY

SOURCE: Saved artifact set
PRIVACY: Hash only
REPORT: Redacted
HARD STOP: Number or entity loss

PROVIDER ELIGIBILITY

3 eligible 1 missing connector

No document text is written to analytics or shared logs.

PROVIDER MATRIX

Candidate decisions

3 candidates / 3 providers

Provider attempts and candidate evidence
Decision	Provider / model	Semantic	Detector	Protected	Cost	Latency	Reason	Freshness
✓ safe selected	Atlas Editatlas-3.2-long · eu-west	0.947pass	3/41 timeout	8/8 passclaims + entities	$0.071	8.42s	All hard gates passed	14m
≈ best observed	Local Editorialmistral-small-3.1 · sto-1	0.919pass	4/4lowest signal	8/8 passclaims + entities	$0.042	4.18s	Voice fit below target	14m
× rejected	Vector Prosevp-rewrite-2 · us-east	0.861hard fail	2/42 unsupported	7/8 failnumber removed	$0.071	12.84s	Number mismatch	22m

POLICY CONSOLE

Resolved production policy

Current v24 · draft v25 · last activated by Pedro D.

Effective cascade

GLOBAL DEFAULTplatform-safe@18Inherited
PRODUCT PACKhumanizerops@94 changes
WORKSPACEagonist-prod@242 overrides
PROJECTariada-productionEU only
API KEYcontent-pipelineBudget cap
RESOLVED POLICYproduction-safe@24Digest 71f…e09

VERSION DIFF

v24 → draft v25

3 changed values

Field	Current v24	Draft v25	Impact
semantic.min_similarity	0.88	0.91	+8.4% rejections
routing.max_attempts	3	2	−19% max cost
privacy.external_regions	eu, uk	eu	1 provider ineligible

! Capability warning Vector Prose becomes ineligible for 14% of current traffic.

! Retry stacking Client retry × route retry may produce four billable calls.

API CONTROL PLANE

Virtual keys and client scopes

Key values are shown once at creation and never stored in this browser.

Project keys

Key	Scope	Policy	Budget	Last used	State
content-pipeline-prodhk_prod_…8F2C	transform:write reports:read	production-safe@24	$340 / $500	18:38 UTC	● ACTIVE
evaluation-cihk_test_…19AD	benchmarks:write	eval-broad@11	$42 / $100	17:04 UTC	● ACTIVE
agency-previewhk_live_…REVOKED	reports:read	client-safe@5	—	12 Jul	× REVOKED

QUICK START

Run the safe route

curl https://api.humanizerops.com/v1/routes/editorial-safe \
  -H "Authorization: Bearer $HOPS_KEY" \
  -H "Idempotency-Key: run_20260714_01" \
  -d @request.json

Effective project: prj_ariada_prod
Default policy: production-safe@24
Region: EU only
Retention: Hash only

Open API reference →

Humaniser engine workbench

Analyze, understand, and revise

Detect AI-generated text

Build a treatment profile, not a magic rewrite

Ordered transformations